SPARK-1128: set hadoop task properties when constructing HadoopRDD

https://spark-project.atlassian.net/browse/SPARK-1128

The task properties are not set when constructing HadoopRDD in current implementation, this may limit the implementation based on

```
mapred.tip.id
mapred.task.id
mapred.task.is.map
mapred.task.partition
mapred.job.id
```

This patch also contains a small fix  in createJobID (SparkHadoopWriter.scala), where the current implementation actually is not using time parameter

Author: CodingCat <zhunansjtu@gmail.com>
Author: Nan Zhu <CodingCat@users.noreply.github.com>

Closes #101 from CodingCat/SPARK-1128 and squashes the following commits:

ed0980f [CodingCat] make SparkHiveHadoopWriter belongs to spark package
5b1ad7d [CodingCat] move SparkHiveHadoopWriter to org.apache.spark package
258f92c [CodingCat] code cleanup
af88939 [CodingCat] update the comments and permission of SparkHadoopWriter
9bd1fe3 [CodingCat] move configuration for jobConf to HadoopRDD
b7bdfa5 [Nan Zhu] style fix
a3153a8 [Nan Zhu] style fix
c3258d2 [CodingCat] set hadoop task properties while using InputFormat